We built a web application through which people can explain the predictions of various classifiers on different datasets using LIME as well as aLIME They needn’t have any expertise in programming or machine learning. We’ve made the entire cumbersome process end to end. The user just needs to upload the dataset in a rectangular format consisting only numerical entries with no entries missing. Then she needs to type in various features and class labels that the dataset comprises of.
After that she can select various parameters like the explainer, classifier, number of top features, etc.
- Clone the repository
- Run
pip3 install -r requirements.txtin the terminal - Run
python3 manage.py runserver - Go to a Web Browser and type
http://localhost:8000/lime/upload
Some sample screenshots of the process are shown in the figure below.
